1. Papers
Risk factors for coronary artery disease in non-insulin
dependent diabetes mellitus: United Kingdom prospective
diabetes study (UKPDS: 23)
R C Turner, H Millns, H A W Neil, I M Stratton, S E Manley, D R Matthews, R R Holman for the
United Kingdom Prospective Diabetes Study Group
Abstract Previous prospective studies of patients with type 2 Diabetes Research
Laboratories,
diabetes mellitus had comparatively few patients and Nuffield
Objective: To evaluate baseline risk factors for cardiovascular end points.5–11 Many of these studies Department of
coronary artery disease in patients with type 2 have not measured the concentration of low density Medicine,
diabetes mellitus. lipoprotein cholesterol, potentially the most important
University of
Oxford, Radcliffe
Design: A stepwise selection procedure, adjusting for
lipid fraction.12 13 Infirmary, Oxford
age and sex, was used in 2693 subjects with complete OX2 6HE
We report a prospective study of white patients
data to determine which risk factors for coronary R C Turner,
with recently diagnosed type 2 diabetes mellitus. After professor of medicine
artery disease should be included in a Cox
entry to the United Kingdom prospective diabetes H Millns,
proportional hazards model.
study14 patients were assessed for baseline risk factors statistician
Subjects: 3055 white patients (mean age 52) with H A W Neil,
after initial treatment by diet for 3 months. The associa-
recently diagnosed type 2 diabetes mellitus and lecturer in clinical
tion of coronary artery disease with baseline risk epidemiology
without evidence of disease related to atheroma.
factors has been assessed irrespective of subsequent I M Stratton,
Median duration of follow up was 7.9 years. 335
treatments. statistician
patients developed coronary artery disease within 10 S E Manley,
years. biochemist
Outcome measures: Angina with confirmatory D R Matthews,
abnormal electrocardiogram; non-fatal and fatal Subjects and methods consultant physician
in diabetes
myocardial infarction. R R Holman,
Patients
Results: Coronary artery disease was significantly reader in medicine
Between 1977 and 1991, 5102 patients aged 25 to 65
associated with increased concentrations of low Correspondence to:
years with type 2 diabetes mellitus based on a fasting
density lipoprotein cholesterol, decreased Professor Turner
plasma glucose concentration > 6 mmol/l on two occa- robert.turner@
concentrations of high density lipoprotein cholesterol, drl.ox.ac.uk
sions were recruited to the study14; 4775 (94%) had fast-
and increased triglyceride concentration,
ing plasma glucose values > 7.0 mmol/l, which is
haemoglobin A1c, systolic blood pressure, fasting BMJ 1998;316:823–8
consistent with the American Diabetic Association’s
plasma glucose concentration, and a history of
definition of diabetes. Of the 7108 patients originally
smoking. The estimated hazard ratios for the upper
referred for entry to the study, 2006 (28%) were
third relative to the lower third were 2.26 (95%
excluded. These were of a similar age and sex and had a
confidence interval 1.70 to 3.00) for low density
similar fasting plasma glucose concentration as those
lipoprotein cholesterol, 0.55 (0.41 to 0.73) for high
patients included in the study. The main reasons for
density lipoprotein cholesterol, 1.52 (1.15 to 2.01) for
haemoglobin A1c, and 1.82 (1.34 to 2.47) for systolic exclusion were myocardial infarction in the previous
blood pressure. The estimated hazard ratio for year, current angina or heart failure, accelerated
smokers was 1.41(1.06 to 1.88). hypertension, proliferative or preproliferative retino-
Conclusion: A quintet of potentially modifiable risk pathy, renal failure with a plasma creatinine concentra-
factors for coronary artery disease exists in patients tion > 175 mol/l, other life threatening disease such as
with type 2 diabetes mellitus. These risk factors are cancer, an illness requiring systemic steroids, an occupa-
increased concentrations of low density lipoprotein tion which precluded insulin treatment, language
cholesterol, decreased concentrations of high density difficulties, or ketonuria > 3 mmol/l suggestive of
lipoprotein cholesterol, raised blood pressure, insulin dependent diabetes mellitus.
hyperglycaemia, and smoking. The study was approved by the ethics committee in
each of the 23 centres: Radcliffe Infirmary, Oxford;
Royal Infirmary, Aberdeen; General Hospital, Bir-
Introduction mingham; St George’s Hospital and Hammersmith
Patients with type 2 diabetes mellitus have a twofold to Hospital, London; City Hospital, Belfast; North
threefold increased incidence of diseases related to Staffordshire Royal Infirmary, Stoke on Trent; Royal
atheroma,1 and those who present in their 40s and 50s Victoria Hospital, Belfast; St Helier Hospital, Car-
have a twofold increased total mortality.2 In the United shalton; Whittington Hospital, London; Norfolk and
Kingdom the incidence of macrovascular complica- Norwich Hospital; Lister Hospital, Stevenage; Ipswich
tions in patients with type 2 diabetes mellitus is twice Hospital; Ninewells Hospital, Dundee; Northampton
that of microvascular disease.3 The greater mortality in Hospital; Torbay Hospital; Peterborough General Hos-
patients with type 2 diabetes mellitus than in the pital; Scarborough Hospital; Derbyshire Royal Infir-
general population cannot be explained only by the mary; Manchester Royal Infirmary; Hope Hospital,
presence of the three classic risk factors for coronary Salford; Leicester General Hospital; Royal Devon and
artery disease—that is, smoking, hypertension, and an Exeter Hospital. All patients gave their informed
increased plasma cholesterol concentration.4 consent to take part in the study.
BMJ VOLUME 316 14 MARCH 1998 823

2. Papers
The initial treatment by diet for 3 months was com- Follow up, identification, and classification of end
pleted by 4178 white patients,14 with a mean loss of points
5 kg body weight, but only 867 (16.9%) were able to Patients were seen every three months in the clinics, and
achieve a near normal fasting plasma glucose concen- any events that were clinically important were noted. To
tration of < 6 mmol/l. Cardiovascular disease was evi- ascertain whether predetermined criteria for the end
dent in 381 (7.5%) patients, of whom 58 (15.2%) had points were attained two independent doctors received
previous myocardial infarction, 144 (37.8%) a definite full information on the patients but without details of
electrocardiographic Q wave abnormality on Minne- treatment.14 Any discrepancies between the two doctors
sota coding, 7 (1.8%) angina, 1 (0.3%) heart failure, 120 were adjudicated by two independent senior doctors. All
end points were coded according to ICD-9 (inter-
(31.5%) intermittent claudication, and 51 (13.4%) a
national classification of disease, 9th revision).15
previous stroke or transient ischaemic attack. Only
Three aggregate end points were evaluated:
2693 (70.9%) of the 3797 datasets could be analysed
coronary artery disease—that is, fatal or non-fatal myo-
for all variables, as biochemical measurements were
cardial infarction or clinical angina—with an abnormal
not undertaken until 1981, and some patients had no
electrocardiogram at rest or after a treadmill test; fatal
valid data for one or more of the other variables. Suffi- or non-fatal myocardial infarction; and fatal myo-
cient data were available from 3055 (80.5%) patients cardial infarction.
for the final Cox model analysis, and 2161 (56.9%)
patients had ophthalmic photographic data available Baseline risk factors assessment
for the assessment of the effect of retinopathy. Height, waist, and hip circumferences were measured,
After the initial treatment diet, patients were and the smoking status and amount of exercise taken
randomly allocated to different treatments according were ascertained by questionnaire.14 Retinopathy was
to the protocol of the United Kingdom prospective assessed by modified Wisconsin grading of four colour
diabetes study.14 This paper does not include any refer- photographs of each eye taken at 30° to the
ence to treatment allocations, actual treatment, or horizontal.14 Blood pressure was recorded as the mean
diabetes control during the 10 years of follow up. of measurements taken 2 and 9 months after diagnosis
with electronic sphygmomanometers. Hypertension
was defined as systolic blood pressure >160 mm Hg or
Table 1 Baseline characteristics in 2693 white patients with no indication of disease diastolic blood pressure >90 mm Hg, or both, or anti-
related to atheroma. Results are means (SD) or geometric mean (1 SD interval) unless hypertensive treatment. After the initial treatment diet,
stated otherwise patients were fasted overnight and the following
Variable Men (n=1564) Women (n=1129) concentrations measured: fasting plasma glucose,
Age (years) 52 (9) 53 (9) haemoglobin A1c, low density lipoprotein cholesterol,
Body mass index (kg/m2) 27.1 (4.7) 29.4 (6.4) high density lipoprotein cholesterol, and insulin.14 16
Waist:hip ratio 0.95 (0.06) 0.87 (0.08)
Systolic blood pressure (mm Hg) 133 (18) 139 (20) Statistical analyses
Diastolic blood pressure (mm Hg) 82 (10) 83 (10) Data are reported as means (SD), geometric means
No (%) of patients with hypertension 508 (32) 506 (45) (1 SD interval), or percentages. Variables for patients
Fasting plasma glucose (mmol/l) 8.3 (2.8) 9.2 (3.0)
included in or excluded from the analyses were
Haemoglobin A1c (%) 6.9 (1.7) 7.4 (1.8)
compared by t tests, 2 tests, or Fisher’s exact tests.
Total cholesterol (mmol/l) 5.2 (1.0) 5.7 (1.2)
Standardised mortality ratio in the patients was cal-
Low density lipoprotein cholesterol (mmol/l) 3.3 (0.9) 3.8 (1.1)
High density lipoprotein cholesterol (mmol/l) 1.04 (0.23) 1.10 (0.24)
culated from the Office of Population Censuses and
Triglyceride (mmol/l) 1.5 (0.9, 2.4) 1.6 (1.0, 2.6)
Surveys death rates for the general population of
Fasting plasma insulin (mU/l) 11.3 (6.6, 19.4) 13.2 (7.7, 22.7) England and Wales for the same calendar period, with
Exercise (No (%) of subjects) stratification by sex and age in periods of five years.17 18
Sedentary 261 (17) 251 (22) Age was categorised as < 50, 50-54, 55-59, or >60
Moderate 496 (32) 440 (39) years. Continuous variables were grouped into thirds.
Active 692 (44) 428 (38) The effect of potential risk factors on the three aggre-
Fit 115 (7) 10 (1) gate end points was assessed by Cox proportional haz-
Smoking (No (%) of subjects) ards models,19 with censoring at 10 years’ follow up.
Never 344 (22) 501 (44) The relation of single risk factors with events after
Ex-smoker 712 (46) 297 (27)
adjustment for age and sex was assessed in 2693
Current 508 (32) 331 (29)
patients with all risk factors measured. Multivariate
selection of risk factors was done by a stepwise proce-
Table 2 Standardised mortality ratios for 5071 patients recently diagnosed with dure after adjustment for age and sex. Estimated
non-insulin dependent diabetes mellitus compared with general population hazard ratios are represented graphically, with 95%
confidence intervals estimated for each group by treat-
Years since No of Standardised
randomisation patients No observed No expected mortality ratio P value ing the relative risks as floating absolute risks so that
Men the appropriate variability for each group is shown.20
0 to <5 2992 153 162 0.94 0.78 Baseline biochemical and blood pressure values
5 to <10 2267 161 118 1.36 <0.001 were corrected for any effects from regression to the
>10 566 42 26 1.62 0.002 mean by examining repeat values at six months in 497
Women patients who had remained on the diet treatment
0 to <5 2079 69 72 0.96 0.64 alone. The effect of a unit increment of risk factors on
5 to <10 1581 84 55 1.52 <0.001
coronary artery disease (1 mmol/l in low density
>10 409 28 12 2.42 <0.0001
lipoprotein cholesterol concentration, 0.1 mmol/l in
824 BMJ VOLUME 316 14 MARCH 1998

3. Papers
Table 3 Relation of potential risk factors to cardiac end points after adjustment for age and sex, in 2693 white patients with non-insulin dependent diabetes
mellitus
Non-fatal or fatal myocardial infarction
Distribution Coronary artery disease (n=280) (n=192) Fatal myocardial infarction (n=79)
Lower Upper Estimated hazard ratios Estimated hazard ratios for Estimated hazard ratios
Variable third third P value for each third* P value each third* P value for each third
Body mass index (kg/m2) 24.8 29.0 0.65 0.083 0.46
Waist:hip ratio 0.87 0.94 0.89 0.50 0.89
Systolic blood pressure (mm Hg) 125 142 0.0032 1 1.52 1.72 0.027 1 1.44 1.70 0.011 1 1.14 2.17
Diastolic blood pressure (mm Hg) 79 87 0.025 1 1.08 1.45 0.0061 1 1.17 1.72 <0.0001 1 0.79 2.09
Hypertension 0.018 1 1.34 0.022 1 1.40 0.008 1 1.83
Fasting plasma glucose (mmol/l) 7.3 9.7 0.016 1 1.31 1.54 0.13 0.017 1 1.83 2.24
Haemoglobin A1c (%) 6.2 7.5 0.0003 1 1.64 1.78 0.01 1 1.47 1.71 0.0099 1 1.09 2.11
Cholesterol (mmol/l) 4.88 5.77 <0.0001 1 1.79 1.93 0.0086 1 1.62 1.67 0.027 1 1.98 2.01
Low density lipoprotein cholesterol 3.02 3.89 <0.0001 1 1.48 2.29 0.0002 1 1.44 2.11 0.0043 1 1.06 2.25
(mmol/l)
High density lipoprotein 0.95 1.15 <0.0001 1 0.87 0.51 0.0085 1 0.84 0.57 0.42
cholesterol (mmol/l)
Triglyceride (mmol/l) 1.22 1.87 <0.0001 1 1.63 1.93 0.011 1 1.40 1.72 0.079
Insulin (mU/l) 9.7 15.6 0.16 0.022 1 1.18 1.63 0.86
Exercise (sedentary, moderate, 0.54 0.044 1 0.73 0.58 0.74† 0.57
active, fit)
Smoking (never smoked, 0.016 1 1.18 1.55 0.015 1 1.27 1.74‡ 0.57
ex-smoker, current smoker)
*Not given for non-significant data. †For the four categories of exercise. ‡For the three categories of smoking.
high density lipoprotein cholesterol concentration, and triglyceride did not differ in subjects according to
10 mm Hg in systolic blood pressure, and 1% in the presence of disease related to atheroma (P > 0.01).
haemoglobin A1c) was estimated by fitting each factor
as a continuous variable in a stepwise selected Cox Standardised mortality ratio
model, leaving other risk factors as categorical During the first five years of the study the standardised
variables and adjusting for regression to the mean. Sta- mortality ratio was not very different from that in the
tistical analyses were performed using sas version 6.1. general population, possibly because patients with life
threatening illnesses were excluded from the study
(table 2). After the first five years of the study patients
Results with type 2 diabetes mellitus had an increased total
Table 1 shows the baseline risk factors assessed for the mortality compared with the general population.
2693 white patients who had no previous indication of
disease related to atheroma and had complete data Relation of baseline risk factors with adjustment
when studied after the initial treatment diet. The men for age and sex
who were entered into the United Kingdom prospective Table 3 shows the relation of potential risk factors,
diabetes study but excluded from this analysis because of stratified by thirds, to coronary artery disease in 280
previous cardiovascular disease were older (mean age 56 patients with end points. Important variables were low
(SD7) years), had higher systolic blood pressure (mean density lipoprotein cholesterol concentration, high
141(20) mm Hg), were more likely to be hypertensive density lipoprotein cholesterol concentration, and also
(46%), and were more likely to be smokers (10% had triglyceride concentration, haemoglobin A1c, systolic
never smoked, 51% were ex-smokers, and 39% were blood pressure, fasting plasma glucose concentration,
current smokers) (P < 0.01 for each). Women who were and smoking; each of these had a positive association
excluded from the study were also older (mean age 56 except high density lipoprotein cholesterol concentra-
(8) years) and had higher systolic blood pressure (mean tion. Similar associations were seen for fatal or
145 (20) mm Hg) (P < 0.01 for each). Baseline non-fatal myocardial infarction (192 patients with end
concentrations of total cholesterol, low density lipopro- points) and fatal myocardial infarction (79 patients
tein cholesterol, high density lipoprotein cholesterol, with end points). Retinopathy was associated with fatal
Table 4 Stepwise selection of risk factors, adjusted for age and sex, in 2693 white patients with non-insulin dependent diabetes
mellitus with dependent variable as time to first event. P values are significance of risk factor after accounting for all other risk factors
in model
Coronary artery disease (n=280) Non-fatal or fatal myocardial infarction (n=192) Fatal myocardial infarction (n=79)
Position in
model Variable P value Variable P value Variable P value
First Low density lipoprotein <0.0001 Low density lipoprotein 0.0022 Diastolic blood pressure 0.0012
cholesterol cholesterol
Second High density lipoprotein 0.0001 Diastolic blood pressure 0.0074 Low density lipoprotein 0.012
cholesterol cholesterol
Third Haemoglobin A1c 0.0022 Smoking 0.025 Haemoglobin A1c 0.024
Fourth Systolic blood pressure 0.0065 High density lipoprotein 0.026
cholesterol
Fifth Smoking 0.056 Haemoglobin A1c 0.053
BMJ VOLUME 316 14 MARCH 1998 825

4. Papers
that were not important were body mass index, waist to
3
hip ratio, exercise, triglyceride concentration, and fast-
Estimated hazard ratios (95% CI)
ing plasma glucose or insulin concentration, and these
were not included in the final model. The exclusion of
triglyceride concentration and not high density
1
lipoprotein cholesterol concentration was not due
solely to imprecision of measurements, since baseline
triglyceride concentration correlated with the values 6
0.4 months later in subjects randomised to, and remaining
40 45 50 55 60 65 70 2.5 3.0 3.5 4.0 4.5 5.0 0.9 1.0 1.1 1.2 1.3
Age (years) Low density lipoprotein High density lipoprotein on, the diet (rs = 0.72 (95% confidence interval 0.68 to
cholesterol (mmol/l) cholesterol (mmol/l) 0.76)). This correlation was greater than the correlation
3 between repeated high density lipoprotein measure-
Estimated hazard ratios (95% CI)
ments (rs = 0.52 (0.45 to 0.58)). Baseline triglyceride
concentration correlated with high density lipoprotein
cholesterol concentration (rs = − 0.27 (P < 0.0001)). No
1 significant interaction term was identified between
high density lipoprotein cholesterol concentration and
low density lipoprotein cholesterol concentration.
0.4
5 6 7 8 9 110 120 130 140 150 160 Never Ex- Current
smokers smokers smokers
Estimated hazard ratios
Haemoglobin A1c (%) Systolic blood pressure (mm Hg)
Table 5 and the figure show the Cox model estimated
Estimated hazard ratios for significant risk factors for coronary artery disease occurring in
335 out of 3055 diabetic patients (expressed as floating absolute risks)
hazard ratios for age and sex and the stepwise selected
variables for coronary artery disease in 3055 patients
(335 with a coronary artery disease). A similar pattern
myocardial infarction (P = 0.005) but not with fatal or of hazard ratios was seen for fatal or non-fatal myo-
non-fatal myocardial infarction (P = 0.124) or with cardial infarction (233 patients with an event) and fatal
coronary artery diseases (P = 0.082). myocardial infarction (103 patients with an event).
Although diastolic blood pressure had a stronger rela-
Stepwise selection of risk factors tion than systolic blood pressure with any myocardial
Table 4 shows the selected risk factors with P values infarction or fatal myocardial infarction, replacement
from the stepwise multivariate Cox models. Factors by systolic blood pressure did not affect the results.
Fitting the risk factors for coronary artery disease as
continuous variables, with allowance for regression to
Table 5 Estimated hazard ratios (95% confidence intervals) for coronary artery disease
in 3055 patients, fitting same explanatory variables for all three dependent variables mean, indicated that for each increment of 1 mmol/l in
low density lipoprotein cholesterol concentration there
Coronary artery Fatal or non-fatal Fatal myocardial
Dependent variable disease myocardial infarction infarction
was a 1.57-fold (95% confidence interval 1.37 to 1.79)
No of patients with event 335 233 103 increased risk of coronary artery disease. For each posi-
Age (years) tive increment of 0.1 mmol/l in high density lipoprotein
<50 1 1 1 cholesterol concentration there was a 0.15-fold (0.08 to
50-54 1.65 (1.18 to 2.31) 1.76 (1.15 to 2.70) 5.56 (2.07 to 14.95) 0.22) decrease in risk, for each increment of 10 mm Hg
55-59 1.78 (1.29 to 2.46) 2.41 (1.62 to 3.57) 8.70 (3.37 to 22.48) in systolic blood pressure a 1.15-fold (1.08 to 1.23)
>60 2.35 (1.71 to 3.23) 2.80 (1.88 to 4.17) 13.86 (5.43 to 35.41) increase, and for each increment of 1% in haemoglobin
Sex A1c a 1.11-fold (1.02 to 1.20) increase in risk.
Women 1 1 1 The retinopathy grading was not significantly
Men 2.12 (1.65 to 2.73) 2.62 (1.91 to 3.58) 4.13 (2.53 to 6.76) related to any of the three aggregate end points when
Low density lipoprotein cholesterol (mmol/l)
added to the multivariate models, including age, sex,
<3.02 1 1 1
and the other risk factors in table 4.
>3.02 to <3.89 1.41 (1.05 to 1.90) 1.41 (1.00 to 2.00) 1.15 (0.67 to 1.97)
>3.89 2.26 (1.70 to 3.00) 2.11 (1.50 to 2.95) 2.32 (1.41 to 3.81)
High density lipoprotein cholesterol (mmol/l) Discussion
<0.95 1 1 1**
>0.95 to <1.15 0.90 (0.71 to 1.15) 0.94 (0.70 to 1.26) 1.16 (0.73 to 1.84) This study shows that in patients with type 2 diabetes
>1.15 0.55 (0.41 to 0.73) 0.65 (0.47 to 0.91) 0.84 (0.51 to 1.39) mellitus increased concentrations of low density
Haemoglobin A1c (%) lipoprotein cholesterol, decreased concentrations of
<6.2 1 1 1 high density lipoprotein cholesterol, hyperglycaemia,
>6.2 to <7.5 1.47 (1.12 to 1.95) 1.26 (0.91 to 1.75) 0.98 (0.58 to 1.65) hypertension, and smoking (all measured on comple-
>7.5 1.52 (1.15 to 2.01) 1.42 (1.03 to 1.98) 1.72 (1.06 to 2.77) tion of a treatment diet after diagnosis), are risk factors
Systolic blood pressure (mm Hg) for coronary artery disease, defined as fatal and
<125 1 1* 1* non-fatal myocardial infarction or angina. Previous
>125 to <142 1.52 (1.12 to 2.06) 1.45 (1.01 to 2.08) 1.22 (0.66 to 2.24) studies have shown inconsistent results, being depend-
>142 1.82 (1.34 to 2.47) 1.76 (1.22 to 2.54) 2.36 (1.33 to 4.18)
ent on univariate analyses in small studies, with few
Smoking
patients having clinical end points. Total cholesterol
Never smoked 1 1 1**
concentration was reported to be a risk factor in
Ex-smoker 1.03 (0.77 to 1.37) 1.08 (0.75 to 1.54) 0.65 (0.39 to 1.09)
Current smoker 1.41 (1.06 to 1.88) 1.58 (1.11 to 2.25) 1.03 (0.62 to 1.70)
some5 7 21 but not other studies,6 8 22 and most had not
measured both low density lipoprotein cholesterol and
*Estimated hazard ratios shown for systolic blood pressure despite diastolic blood pressure being included
in stepwise selection. high density lipoprotein cholesterol concentrations.
**Variable not included in stepwise selection model. Hyperglycaemia was similarly reported as a risk factor
826 BMJ VOLUME 316 14 MARCH 1998

5. Papers
in some5 6 8 11 21 22 but not other studies,7 and hyperten- pathogenic factor for coronary artery disease the
sion similarly in some5 22 but not other studies.6–8 11 21 increased risk would be expected to be proportional to
The present study shows that patients with type 2 the degree of hyperglycaemia.34 The study showed an
diabetes mellitus have the same risk factors for increased risk of 11% for each increment of 1% in
coronary artery disease as the general population.23 haemoglobin A1c, similar to the 10% increase in mor-
This study confirms that patients with non-insulin tality from ischaemic heart disease for an increment of
diabetes mellitus have an increased total mortality 1% in haemoglobin A1c reported in Wisconsin.35
compared with the general population, although this Blood pressure—Increased blood pressure was also a
was not apparent in the initial 5 years, probably major risk factor for coronary artery disease, with a 15%
because diabetic patients with life threatening illness increased risk for an increase in systolic blood pressure
were excluded from the United Kingdom prospective of 10 mm Hg, which was similar to that reported in the
diabetes study. general population.36 Increased blood pressure was a
major risk factor for fatal myocardial infarction. This
Major risk factors could be because hypertension is a major additional
Low density lipoprotein and total cholesterol—An burden to the heart when myocardial infarction ensues.
increased concentration of low density lipoprotein The hypertension in diabetes study in 1148 patients in a
cholesterol or total cholesterol at baseline was a major factorial design is evaluating whether strict blood
risk factor for coronary artery disease. This is similar to pressure control will prevent complications.37
the general population.13 24 25 Increased concentrations Retinopathy—Retinopathy at diagnosis was not a risk
of low density lipoprotein cholesterol may be more factor for cardiovascular disease in a multivariate
pathogenic in patients with type 2 diabetes mellitus than analysis, although retinopathy was a risk factor for fatal
in non-diabetic patients because of the presence of small myocardial infarction when only age and sex were
dense low density lipoprotein cholesterol particles26 and adjusted for. Since both retinopathy38 and micro-
oxidation of glycated low density lipoprotein albuminuria39 are associated with hyperglycaemia and
cholesterol.27 The 1.57 increased risk for an increment of hypertension, which are also risk factors for coronary
1 mmol/l in low density lipoprotein cholesterol concen- artery disease, the previously described association of
tration equates to a 36% risk reduction for a decrement retinopathy and microalbuminuria with subsequent
of 1 mmol/l , similar to the 31% risk reduction achieved cardiovascular mortality might reflect the longstanding
with a 3-hydroxy-3-methylglutaryl coenzyme A reduct- hypertension and hyperglycaemia that induced both
ase inhibitor in men with hypercholesterolaemia.13 The macrovascular and microvascular disease.
subgroup analysis of the simvastatin study28 showed that
the diabetic patients had similar protection to that of
Risk factors in type 2 diabetes mellitus
non-diabetic patients.25 A decreased concentration of
Risk factors for development of coronary artery disease
high density lipoprotein cholesterol was an independent
in the general population may not apply once diabetes
risk factor for coronary artery disease. The 15% decrease
has developed. Obesity and central obesity,40 decreased
in the risk of coronary artery disease associated with a
physical activity,41 and raised insulin concentrations42
0.1 mmol/l increment in high density lipoprotein
provide an increased risk for cardiovascular disease, but
cholesterol concentration is compatible with the 8-12%
in patients with type 2 diabetes mellitus we found that
reduction reported from prospective American
none of these were major risk factors. These variables
studies.29 Triglyceride concentration was a risk factor for
are also risk factors for diabetes,43–45 but this study
coronary artery disease after adjustment for age and sex,
indicates that once diabetes has developed, hyperten-
but it was not an independent risk factor when the other
sion, increased concentrations of low density lipoprotein
variables were included in the model. This is in accord
or decreased concentrations of high density lipoprotein
with other studies, possibly because of the greater
cholesterol and hyperglycaemia measured at baseline
biological variability of triglyceride than high density
are greater risk factors for coronary artery disease than
lipoprotein cholesterol measurements.30 However, we
these precipitating factors.
found over 6 months that the concentration of high
Syndrome X, the association of raised concentra-
density lipoprotein cholesterol was more variable than
tions of glucose, insulin, and triglyceride, decreased con-
that of triglyceride, possibly because of less precision
centrations of high density lipoprotein cholesterol, and
with the assay (coefficient of variation 6% v 2%) and
increased blood pressure, describes a combination of
because patients were receiving dietary advice and had a
previously reported risk factors for coronary artery dis-
more uniform dietary intake than in the general
ease.46 In the general population the combination of
population. As control of plasma triglyceride and high
upper body obesity, glucose intolerance, hypertriglyceri-
density lipoprotein cholesterol concentrations is inter-
daemia, and hypertension has been termed the deadly
linked through lipoprotein lipase and hepatic lipase
quartet.47 However, a quintet of increased concentrations
activities, it may not be feasible to separate the contribu-
of low density lipoprotein and decreased concentrations
tions of triglyceride and high density lipoprotein
of high density lipoprotein cholesterol, hypertension,
cholesterol to coronary artery disease. Postprandial tri-
hyperglycaemia, and smoking is probably more relevant
glyceride values may have an additional atherogenic role
in patients with type 2 diabetes mellitus.
to the fasting values that were measured.31
Haemoglobin A1c—There was an increase in risk of The cooperation of the patients and many NHS and non-NHS
coronary artery disease with haemoglobin A1c of staff at the centres is much appreciated. We thank Professor Eva
> 6.2%, the upper range of normal values, in accord Kohner, Mr Steve Aldington, Ms Ivy Samuel, and Mrs Caroline
Wood for help with the manuscript.
with other studies which suggest that glycaemia above Contributors: RCT and RRH coordinated the study, HM
the normal range gives an increased risk for macro- and IMS carried out the statistical analyses, HAWN provided
vascular disease.32 33 If glycation of proteins was a major epidemiological advice, DRM coordinated the assessment of
BMJ VOLUME 316 14 MARCH 1998 827

6. Papers
15 World Health Organisation. International classification of procedures in
Key messages medicine. Geneva: World Health Organisation, 1978.
16 United Kingdom Prospective Diabetes Study Group. UKPDS XI.
Biochemical risk factors in type 2 diabetic patients at diagnosis compared
x Coronary artery disease is the major cause of mortality in patients with age-matched normal subjects. Diabetic Med 1994;11:534-4.
with type 2 diabetes mellitus 17 Population Censuses and Surveys Office. Mortality statistics: review of the
registrar general on deaths in England and Wales. London: HMSO, 1994.
x Patients without evidence of disease related to atheroma at diagnosis 18 Coleman MP, Herman C, Douglas A. Person-years (PYRS): a FORTRAN
of type 2 diabetes mellitus had an increased standardised mortality program for cohort study analysis. Lyons: International Agency for Research
on Cancer, 1989.
ratio compared with the population of the United Kingdom 19 Cox DR. Regression models and life-tables. J R Stat Soc Series B
1972;34:187-201.
x 11% of patients in this study had a myocardial infarction or 20 Easton DF, Peto J, Babiker AG. Floating absolute risk: an alternative to
developed angina over a median of 8 years’ follow up relative risk in survival and case-control analysis avoiding an arbitrary
reference group. Stat Med 1991;10:1025-35.
x The potentially modifiable risk factors for coronary artery disease 21 Lehto S, Ronnemaa T, Haffner SM, Pyörälä K, Kallio V, Laakso M. Dys-
were increased concentrations of low density lipoprotein cholesterol, lipidemia and hyperglycemia predict coronary heart disease events in
middle-aged patients with NIDDM. Diabetes 1997;48:1354-9.
decreased concentrations of high density lipoprotein cholesterol, 22 Walters DP, Gatling W, Houston AC, Mullee MA, Julious SA, Hill RD.
hypertension, hyperglycaemia, and smoking; these are also risk Mortality in diabetic subjects: an eleven-year follow-up of a community-
based population. Diabetic Med 1994;11:968-73.
factors for coronary artery disease in the general population 23 Kannel WB, Wilson PW. An update on coronary risk factors. Med Clin
North Am 1995;79:951-71.
x Evidence is needed on whether modifying these risk factors will 24 Law MR, Wald NJ, Thompson SG. By how much and how quickly does
reduce coronary artery disease in patients with type 2 diabetes reduction in serum cholesterol concentration lower risk of ischaemic
heart disease? BMJ 1994;308:367-72.
mellitus 25 Scandinavian Simvastatin Survival Study Group. Randomised trial of
cholesterol lowering in 4444 patients with coronary heart disease: the
Scandinavian Simvastatin Survival Study (4S). Lancet 1994;344:1383-9.
clinical end points, and SEM provided biochemical advice. All 26 Austin MA, Breslow JL, Hennekens CH, Buring JE, Willett WC, Krauss
authors wrote the paper. RCT will act as guarantor of the study. PM. Low-density lipoprotein subclass patterns and risk of myocardial in-
farction. JAMA 1988;260:1917-21.
Funding: This study received grants from the Medical 27 Kawamura M, Heinecke JW, Chait A. Pathophysiological concentrations
Research Council; British Diabetic Association; the Department of glucose promote oxidative modification of low density lipoprotein by a
of Health; the National Eye Institute and the National Institute superoxide-dependent pathway. J Clin Invest 1994;94:771-8.
of Digestive, Diabetes and Kidney Disease in the National Insti- 28 Pyörälä K, Pedersen TR, Kjekshus J. The effect of cholesterol lowering
tutes of Health, United States; the British Heart Foundation; the with simvastatin on coronary artery diseases in diabetic patients with cor-
nary heart disease. Diabetes 1995;44(suppl):35A.
Health Promotion Research Trust; Charles Wolfson Charitable
29 Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD, Castelli WP, Knoke
Trust; the Alan and Babette Sainsbury Trust; the Clothworkers’ JD, et al. High-density lipoprotein cholesterol and cardiovascular disease.
Foundation; Oxford University Medical Research Fund Four prospective American studies. Circulation 1989;79:8-15.
Committee; and pharmaceutical companies, including Novo- 30 Schectman G, Sasse E. Variability of lipid measurements: relevance for
Nordisk, Bayer, Bristol Myers Squibb, Hoechst, Lilly, Lipha, and the clinician. Clin Chem 1993;39:1495-1503.
31 Stensvold I, Tverdal A, Urdal P, Graff-Iversen S. Non-fasting serum
Farmitalia Carlo Erba. Additional funding came from
triglyceride concentration and mortality from coronary heart disease and
Boehringer Mannheim, Becton Dickinson, Owen Mumford, any cause in middle aged Norwegian women. BMJ 1993;307:1318-22.
Securicor, Kodak, and Cortecs Diagnostics. 32 Jarrett RJ, Keen H. Hyperglycaemia and diabetes mellitus. Lancet
Conflict of interest: None. 1976;2:1009-12.
33 Folsom AR, Szklo M, Stevens J, Liao F, Smith R, Eckfeldt J. A prospective
1 Garcia MJ, McNamara PM, Gordon T, Kannell WB. Morbidity and study of coronary heart disease in relation to fasting insulin, glucose, and
mortality in diabetics in the Framingham population. Sixteen year diabetes. Diabetes Care 1997;20:935-42.
follow-up. Diabetes 1974;23:105-11. 34 Brownlee M. Glycation products and the pathogenesis of diabetic
2 Panzram G. Mortality and survival in type 2 (non-insulin-dependent) complications. Diabetes Care 1992;15:1835-43.
diabetes mellitus. Diabetologia 1987;30:123-31. 35 Klein R. Hyperglycemia and microvascular and macrovascular disease in
3 United Kingdom Prospective Diabetes Study Group. UKPDS 17. A nine- diabetes. Diabetes Care 1995;18:258-68.
year update of a randomized, controlled trial on the effect of improved 36 MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, et al. Blood
metabolic control on complications in non-insulin-dependent diabetes pressure, stroke, and coronary heart disease. Part 1. Prolonged
mellitus. Ann Intern Med 1996;124:136-45. differences in blood pressure: prospective observational studies corrected
4 Stamler J, Vaccaro O, Neaton JD, Wentworth D. Diabetes, other risk for the regression dilution bias. Lancet 1990;335:765-74.
factors, and 12 year cardiovascular mortality for men screened in the 37 Hypertension in Diabetes Study Group. HDS 4: therapeutic require-
Multiple Risk Factor Intervention Trial. Diabetes Care 1993;16:434-44. ments to maintain tight blood pressure control. Diabetologia
5 Gall MA, Borch-Johnsen K, Hougaard P, Nielsen FS, Parving HH. 1996;39:1554-61.
Albuminuria and poor glycaemic control predict mortality in NIDDM. 38 Stratton IM, Kohner E, Aldington S, Matthews DR, Turner RC.
Diabetes 1995;44:1303-9. Prevalence of diabetic retinopathy at diagnosis of non-insulin dependent
6 Kuusisto J, Mykkänen L, Pyörälä K, Laakso M. NIDDM and its metabolic diabetes in 2964 white Caucasian patients and association with hyperten-
control predict coronary heart disease in elderly subjects. Diabetes sion, hyperglycaemia and impaired beta-cell function. Diabetes
1994;43:960-7. 1995;43:S434.
7 Fontbonne A, Eschwège E, Cambien F, Richard JL, Ducimetière P, Thibult 39 United Kingdom Prospective Diabetes Study Group. UKPDS X. Urinary
N, et al. Hypertriglyceridaemia as a risk factor of coronary heart disease albumin excretion over 3 years in diet-treated type 2 (non-insulin-
mortality in subjects with impaired glucose tolerance or diabetes. Results dependent) diabetic patients, and association with hypertension, hyper-
from the 11-year follow-up of the Paris Prospective Study. Diabetologia glycaemia and hypertriglyceridaemia. Diabetologia 1993;36:1021-9.
1989;32:300-4. 40 Byers T. Body weight and mortality. N Engl J Med 1995;333:723-4.
8 Uusitupa MI, Niskanen LK, Siitonen O, Voutilainen E, Pyörälä K. Ten-year 41 Manson JE, Willett WC, Stampfer MJ, Colditz GA, Hunter DJ, Hankinson
cardiovascular mortality in relation to risk factors and abnormalities in SE, et al. Body weight and mortality among women. N Engl J Med
lipoprotein composition in type 2 (non-insulin-dependent) diabetic and 1995;333:677-85.
non-diabetic subjects. Diabetologia 1993;36:1175-84. 42 Desprès JP, Lamarche B, Mauriège P, Cantin B, Dagenais GR, Moorjani S,
9 Butler WJ, Ostrander LD, Carman WJ, Lamphiear DE. Mortality from et al. Hyperinsulinemia as an independent risk factor for ischemic heart
coronary heart disease in the Tecumseh study. Long-term effect of disease. N Engl J Med 1996;334:952-7.
diabetes mellitus, glucose tolerance and other risk factors. Am J Epidemiol 43 Modan M, Karasik A, Halkin H, Fuchs Z, Lusky A, Shitrit A, et al. Effect of
1985;121:541-7. past and concurrent body mass index on prevalence of glucose
10 Heyden S, Heiss G, Bartel AG, Hames CG. Sex differences in coronary intolerance and type 2 (non-insulin-dependent) diabetes and on insulin
mortality among diabetics in Evans County, Georgia. J Chronic Dis response. The Israel study of glucose intolerance, obesity and
1980;33:265-73. hypertension. Diabetologia 1986;29:82-9.
11 Laakso M, Lehto S, Penttila I, Pyörälä K. Lipids and lipoproteins predict- 44 Desprès JP, Moorjani S, Lupien PJ, Tremblay A, Nadeau A, Bouchard C.
ing coronary heart disease mortality and morbidity in patients with non- Regional distribution of body fat, plasma lipoproteins, and cardiovascular
insulin-dependent diabetes. Circulation 1993;88:1421-30. disease. Arteriosclerosis 1990;10:497-511.
12 Law MR, Wald NJ, Wu T, Hackshaw A, Bailey A. Systematic underestima- 45 Manson JE, Rimm EB, Stampfer MJ, Colditz GA, Willett WC, Krolewski
tion of association between serum cholesterol concentration and ischae- AS, et al. Physical activity and incidence of non-insulin-dependent
mic heart disease in observational studies: data from the BUPA study. diabetes mellitus in women. Lancet 1991;338:774-8.
BMJ 1994;308:363-6. 46 Reaven GM. Role of insulin resistance in human disease. Diabetes
13 Shepherd J, Cobbe SM, Ford I, Isles CG, Lorimer AR, MacFarlane PW, et 1988;37:1595-1607.
al. Prevention of coronary heart disease with pravastatin in men with 47 Kaplan NM. The deadly quartet. Upper-body obesity, glucose intolerance,
hypercholesterolemia. West of Scotland Coronary Prevention Study hypertriglyceridemia, and hypertension. Arch Intern Med 1989;149:
Group. N Engl J Med 1995;333:1301-7. 1514-20.
14 United Kingdom Prospective Diabetes Study Group. UKPDS VIII. Study
design, progress and performance. Diabetologia 1991;34:877-90. (Accepted 24 October 1997)
828 BMJ VOLUME 316 14 MARCH 1998